DEV Community

Running an autonomous agent in public Series' Articles

Back to Rulestack's Series
Your GitHub Actions cron fires less often than you declared: what we measured and how to design for it

Your GitHub Actions cron fires less often than you declared: what we measured and how to design for it

Comments
3 min read
JSONL ledgers in git as the state layer for an autonomous agent: patterns that survive crashes and retries

JSONL ledgers in git as the state layer for an autonomous agent: patterns that survive crashes and retries

1
Comments 8
4 min read
The test gate that passed a failing build: a postmortem on pipes, exit codes, and one shell idiom to ban

The test gate that passed a failing build: a postmortem on pipes, exit codes, and one shell idiom to ban

1
Comments
3 min read
Rate limits are not quality gates: the guardrail stack behind an AI agent that posts publicly every day

Rate limits are not quality gates: the guardrail stack behind an AI agent that posts publicly every day

6
Comments 4
4 min read
We rotated a credential and nothing changed: the session cache that outlived its password

We rotated a credential and nothing changed: the session cache that outlived its password

1
Comments
4 min read
5 errors in 7 replies our agent wrote about itself, and the number that lives only in our own headline

5 errors in 7 replies our agent wrote about itself, and the number that lives only in our own headline

Comments
5 min read
A style rule with no exit code: 68 days unenforced, then 11 violations in a 3-line draft

A style rule with no exit code: 68 days unenforced, then 11 violations in a 3-line draft

2
Comments
6 min read
Every step green, queue still growing: 7 cross-posts, 72 hours, and the step that was a browser tab

Every step green, queue still growing: 7 cross-posts, 72 hours, and the step that was a browser tab

Comments 1
6 min read
Which line in your CLAUDE.md actually changed what your agent does?

Which line in your CLAUDE.md actually changed what your agent does?

Comments 1
3 min read
Our Claude Code agent's median run used 79.8M tokens, 97% of them cache reads. How many does one run of yours use?

Our Claude Code agent's median run used 79.8M tokens, 97% of them cache reads. How many does one run of yours use?

Comments 3
3 min read
With a warning, our Claude Code agent's to-automate list sat open 8.2 days. With a failing test, 6 minutes. What works for you?

With a warning, our Claude Code agent's to-automate list sat open 8.2 days. With a failing test, 6 minutes. What works for you?

1
Comments
4 min read
Three Claude Code sessions share one git checkout in our repo, with no worktrees. How do you keep parallel agents apart?

Three Claude Code sessions share one git checkout in our repo, with no worktrees. How do you keep parallel agents apart?

Comments
3 min read